Comparative Morphophysiological Analyses and Molecular ...

ORIGINAL RESEARCHpublished: 05 January 2016

doi: 10.3389/fpls.2015.01184

Frontiers in Plant Science | www.frontiersin.org 1 January 2016 | Volume 6 | Article 1184

Edited by:

Matthias Wissuwa,

Japan International Research Center

for Agricultural Sciences, Japan

Reviewed by:

Guohua Xu,

Nanjing Agricultural University, China

Huixia Shou,

Zhejiang University, China

*Correspondence:

Jitender Giri

[email protected]

†These authors have contributed

equally to this work.

Specialty section:

This article was submitted to

Plant Nutrition,

a section of the journal

Frontiers in Plant Science

Received: 15 September 2015

Accepted: 10 December 2015

Published: 05 January 2016

Citation:

Mehra P, Pandey BK and Giri J (2016)

Comparative Morphophysiological

Analyses and Molecular Profiling

Reveal Pi-Efficient Strategies of a

Traditional Rice Genotype.

Front. Plant Sci. 6:1184.

doi: 10.3389/fpls.2015.01184

Comparative MorphophysiologicalAnalyses and Molecular ProfilingReveal Pi-Efficient Strategies of aTraditional Rice GenotypePoonam Mehra †, Bipin K. Pandey † and Jitender Giri *

Plant Nutrition Laboratory, National Institute of Plant Genome Research, New Delhi, India

Phosphate (Pi) deficiency severely affects crop yield. Modern high yielding rice genotypes

are sensitive to Pi deficiency whereas traditional rice genotypes are naturally compatible

with low Pi ecosystems. However, the underlying molecular mechanisms for low Pi

tolerance in traditional genotypes remain largely elusive. To delineate the molecular

mechanisms for low Pi tolerance, two contrasting rice genotypes, Dular (low Pi tolerant),

and PB1 (low Pi sensitive), have been selected. Comparative morphophysiological,

global transcriptome and lipidome analyses of root and shoot tissues of both genotypes

grown under Pi deficient and sufficient conditions revealed potential low Pi tolerance

mechanisms of the traditional genotype. Most of the genes associated with enhanced

internal Pi utilization (phospholipid remobilization) and modulation of root system

architecture (RSA) were highly induced in the traditional rice genotype, Dular. Higher

reserves of phospholipids and greater accumulation of galactolipids under low Pi in Dular

indicated it has more efficient Pi utilization. Furthermore, Dular also maintained greater

root growth than PB1 under low Pi, resulting in larger root surface area due to increased

lateral root density and root hair length. Genes involved in enhanced low Pi tolerance of

the traditional genotype can be exploited to improve the low Pi tolerance of modern high

yielding rice cultivars.

Keywords: metabolic flexibility, microarray, lipidomics, root system architecture (RSA), phosphate

INTRODUCTION

Phosphorus (P) is a critical element for plant growth and development. It is an essential componentof nucleic acids, membrane lipids, and regulates many vital plant physiological processes likephotosynthesis and respiration. Most of the natural soil P exists in the form of organic compoundsor sparingly soluble cationic complexes. As a result, phosphate (Pi), an inorganic bioavailableform of phosphorus, is a limiting factor for ∼67% of the world’s cultivable soils (Gilbert, 2009).Modern agriculture relies intensively on high input of Pi-fertilizers to compensate for limited soilPi. However, it has been predicted that at the current rate of extraction, global P reserves of rockphosphate will be depleted soon (www.ifdc.org). In addition, application of Pi fertilizers is notfavored economically and environmentally. Ironically, less than 20% of the applied Pi is absorbedby plants whilst the remainder forms insoluble complexes and also runs-off into water bodies (Haand Tran, 2013).

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

Several plant adaptive responses to low Pi have been described.These include modulation of Root System Architecture (RSA)for increasing topsoil foraging (Lynch, 2011; Pandey et al.,2013), enhancing the activity of Pi transporters for its uptake(Ai et al., 2009), secretion of organic acids and phosphatasesto mobilize the Pi from soil organic/inorganic matter (Wanget al., 2011) and production of ribonucleases/lipases to remobilizethe cellular Pi (Raghothama, 1999). Membranes in the form ofphospholipids contain 15–20% of total organic P in a cell (Poirieret al., 1991). Under P starvation, membrane phospholipidsare hydrolyzed, galactolipids which do not contain Pi aresynthesized, and Pi is released as adaptive strategy (Nakamura,2013). These adaptive responses are orchestrated by complexmolecular networks involving several genes (Ha and Tran, 2013;Pant et al., 2015). A multi-component molecular network of Piscavenging systems, the Pho regulon, has been studied widely inplants (Reviewed in López-Arredondo et al., 2014). This systememploys transcription factors, ubiquitin ligases, miRNAs andseveral downstream genes to regulate the Pi homeostasis in plants(Rouached et al., 2010).

Cultivated rice genotypes can be classified as “Pi responsive”(higher yield under high Pi) and “Pi efficient” (yield protectionunder low Pi) (Gerloff, 1977). Most of the modern rice cultivarsare “Pi responsive” as they had been developed and selectedon soils supplemented with Pi fertilizers. These genotypespossess shallow root systems; well-adapted for enhanced Piacquisition from the topsoil under high Pi conditions (Wanget al., 2010; Rose et al., 2011). They show efficient partitioningof photosynthates toward economic yield, contributing to theirhigher harvest index (HI). However, under low Pi conditionsthesemodern genotypes exhibit severe yield losses and are of littlevalue. In contrast, naturally existing low Pi tolerant genotypesincluding landraces and naturally inbred traditional cultivarshave been cultivated on Pi poor soils for very long time and thesegenotypes possess the genetic and phenotypic competence towithstand low Pi conditions (Wissuwa and Ae, 2001). Therefore,traditional genotypes can be an excellent resource of genes toimprove the low Pi tolerance of high yielding modern ricegenotypes.

Global transcriptome analysis in response to Pi deficiency hasbeen performed in Arabidopsis (Misson et al., 2005; Müller et al.,2007), rice (Wasaki et al., 2003, 2006; Pariasca-Tanaka et al., 2009;Li et al., 2010; Dai et al., 2012; Park et al., 2012; Cai et al., 2013;Oono et al., 2013; Secco et al., 2013), tomato (Wang et al., 2002),bean (Hernández et al., 2007), maize (Calderon-Vazquez et al.,2008), and mustard (Hammond et al., 2005). These studies havebeen proven extremely effective in unraveling many novel lowPi responsive genes. However, a comprehensive picture of low Pitolerance mechanisms in naturally tolerant traditional genotypesis missing.

In the present study, we have profiled a modern lowPi sensitive rice genotype (PB1) and a traditional tolerantgenotype (Dular) to investigate their differential behavior underlow Pi. Our comparative morphophysiological, transcriptomeand lipidome analyses of these genotypes revealed thatthe low Pi tolerance of Dular is likely due to efficientinternal Pi remobilization and maintenance of root growth

for Pi uptake under low Pi conditions. We also report thechanges in phospholipid/galactolipids accumulation and theircorresponding genes under low Pi. These Pi-efficient strategiesof traditional genotypes can be exploited to improve the low Pitolerance of high yielding modern rice genotypes.

MATERIALS AND METHODS

Plant Material and Growth ConditionsSeeds of PB1 and Dular were surface-sterilized by 0.1% mercuricchloride for 15min and thereafter, washed with sterile water five-times and germinated on wet filter paper for 2 days. Uniformlygerminated seedlings were transferred to Pi sufficient (320µM)and Pi deficient (1µM) nutrient media (Yoshida et al., 1976)with iron supplemented as FeNaEDTA. Seedlings were grownin growth chamber with 16 h day (30◦C)/8 h night (28◦C)photoperiod, 250–300µM photons/m2/sec photon density and70% relative humidity. Containers filled with 15 liters nutrientsolution were used to grow 30 seedlings per genotype perbiological replicate. The nutrient solution (pH 5.5) was changedevery 24 h. After 15 days, root and shoot tissues were harvestedseparately and immediately frozen in liquid nitrogen for furtheranalyses. Soluble Pi estimation in roots and shoots and In-GelAPase assay were performed as described (Jain et al., 2007; Wanget al., 2011). Root and shoot lengths were measured manuallyusing a ruler.

Analysis of Lateral Roots and Root AngleSeedlings were grown under low and sufficient Pimedia in asepticconditions using MS media with 0.2% phytagel for 15 days. Forlateral root analysis, 15-day-old roots were imaged and lateralroot length and density on primary roots were calculated usingImageJ 1.46r (http://imagej.nih.gov/ij). For the root angle study,seeds were transferred to the center of a basket containing soilritewhich was pre-treated with 1M HCl and repeatedly washed withMilli Q water. Dried soilrite was filled in a basket supported on apot containing Yoshidamedia (+Pi and−Pi), pH 5.0−5.5.Mediawere refreshed after every 24 h. After 30 days, the numbers ofroots emerging from bottom and sides of basket were countedmanually to calculate the root angle.

Extraction of Total RNATotal RNA was extracted from root and shoot tissues of 15-days-old seedlings using RNeasy Mini Kit (Qiagen) according tothe manufacturer’s instruction. Tissues from three seedlings werepooled together for isolation of a sufficient amount of total RNA.Purity and integrity of RNA was determined using Bioanalyzer(2100 Agilent technologies). RNA samples having RIN (RNAIntegrity Number) value above 9.5 in root and 8 in shoot wereused for microarray analysis.

Affymetrix Genechip Hybridization,Washing, and ScanningThe differential gene expression in each tissue/genotype grownunder low and sufficient Pi conditions was analyzed inthree independent biological replicates using Affymetrix rice

Frontiers in Plant Science | www.frontiersin.org 2 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

genome array (57 K) GeneChip R© according to manufacturer’sprotocol. Washing and scanning was carried out at AffymetrixGeneChip R© Fluidics Station 450 and GeneChip R© Scanner3000 7G, respectively. Probe cel intensities for each arraywere retrieved in .cel file format by Affymetrix R© GeneChip R©

Command Console R© Software (AGCC).

Microarray Data AnalysisFor microarray analysis, data from 24 .cel files representingroot and shoot tissues of PB1 and Dular, raised under Pideficient and sufficient conditions in three biological replicateswere imported into “GeneSpring” software (Agilent TechnologiesInc.). The normalization and probe intensities summarizationwere performed by GC-RMA. Three biological replicates ofeach sample, showing correlation coefficient value of ≥ 0.97were considered for downstream analysis (Supplementary Table1). To identify the “significantly expressed” probe sets, aThree-way ANOVA analysis with the following model wasperformed to dissect out the independent effects of all three mainvariables (Genotype, G; Tissue, T; and Phosphate, Pi) and theirinteractions:

Yijkt = µ + Ti + Gj + Pik + (Ti × Gj)+ (Ti × Pik)

+ (Gj × Pik)+ (Ti × Gj × Pik)+ eijkt (1)

Yijkt denotes random variable giving the response for observationt of the treatment at level i, j, k of Tissue, Genotype and Phosphatewhereas eijkt is independent random variable. To decrease thenumber of false positives, Benjamini Hochberg correction wasapplied at p-value cut-off of ≤ 0.05. All subsequent analysis wascarried out as described (Deveshwar et al., 2011).

Corresponding gene IDs for the final dataset were obtainedfrom the Rice Oligonucleotide Array Database (http://www.ricearray.org/), KOME (http://cdna01.dna.affrc.go.jp/cDNA) and NCBI (http://www.ncbi.nlm.nih.gov). Probe setsshowing ≥2-fold changes (FC) in expression under Pi deficiencyin relation to their respective control tissue, were consideredas “differentially expressed” (−Pi/+Pi). For evaluating relativegenotypic effects (Dular/PB1), relative gene expression in Dularwas calculated using PB1 as control. Finally, all significantlyexpressed genes were assigned annotations and GO terms usingRice Genome Annotation Database (http://rice.plantbiology.msu.edu/index.shtml). Differentially expressed genes were thensearched in KEGG (http://www.genome.jp/kegg/pathway.html)and RiceCyc (http://pathway.gramene.org/gramene/ricecyc.shtml) for assigning their putative roles in cellular metabolism.Heat maps were generated in MeV (Multi Experiment Viewer)software, version 4.6.0. on log2 transformed FC and hierarchicalclustering was done using distance matrix Pearson correlation.

Quantitative real time PCR (qRT-PCR) was performed inthree replicates as described earlier (Deveshwar et al., 2011).Primers used for the qRT-PCR are listed in SupplementaryTable 2.

Lipid Profiling by LC-QTOFPlant lipids were extracted from 50mg lyophilised samplesof 15-day-old seedlings as described by Matyash et al.

(2008). The Lipidome data were acquired using an Agilent6530 QTOF for positive ion analysis and Agilent 6550QTOF Mass spectrometer for negative ion analysis (employingAgilent jet stream thermal focussing technology). Raw datawere processed by Agilent’s Mass Hunter Qual software tofind peaks. Peaks were then imported into Mass ProfilerProfessional for peak alignments and filtering. To annotatelipids, MSMS files were used to query the Lipid Blastlibrary.

Identification of DNA PolymorphismsSNPs and InDels in differentially regulated PSR genes wereidentified from whole genome resequencing data of bothgenotypes (Mehra et al., 2015). For identification of SNPs inpromoters of key PSR genes, 2 Kb upstream promoter regionswere analyzed.

RESULTS

Comparative Morphophysiological StudyRevealed Greater Potential of Dular toWithstand Low Pi ConditionsWe have recently shown greater biomass accumulation androot hair growth under low Pi of Dular as compared to PB1,which are low Pi tolerant and sensitive genotypes, respectively(Mehra et al., 2015). Here, we analyzed the effects of genotype,treatment and their interaction on growth of 15-days oldseedlings under low Pi. Reduction in shoot length was morepronounced in PB1 as compared to Dular under low Pi. However,root length in Dular increased by 33% while PB1 showed an11% decrease in root length under low Pi (Figure 1A, Table 1).Factorial analysis showed that all the morphological traits weresignificantly influenced by the low Pi treatment followed bythe effect of genotype (Table 1). Interestingly, both genotypesshowed a similar increase in root to shoot biomass ratio underPi deficiency (Figure 1B). Our Two way ANOVA analysis ofgrowth parameters showed that these traits are influencedsignificantly by the Pi treatment in 15-day-old seedlings (p <

0.05). Intriguingly, Dular has relatively lower soluble Pi contentthan PB1 in both root and shoot tissues under sufficient Pi.However, under low Pi, Dular showed higher Pi accumulationthan PB1 (Table 1).

Dular Root System Architecture Seems tobe Better Adapted for Low Pi ToleranceRSA analysis revealed greater lateral root length anddensity in Dular than PB1 which was further increasedunder low Pi. Interestingly, this increase was higher inDular as compared to PB1 (Figures 1C–E). Our analysisfurther showed shallower root growth of PB1 relative toDular, under both low and sufficient Pi (Figures 1F–H).Interestingly, root spread behavior of both genotypesdid not alter significantly under Pi deficiency. However,the total number of roots decreased under low Pi in PB1(Figures 1F–H).

Frontiers in Plant Science | www.frontiersin.org 3 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

FIGURE 1 | Differential plant growth under Pi deficiency (A) Seedling growth of hydroponically grown Dular and PB1 raised under Pi sufficient (320µM)

and Pi deficient (1µM) conditions in hydroponics after 15-days and (B) two-month. (C) Lateral root phenotype; (D) length and (E) density of PB1 and Dular

under +Pi and −Pi. (F) Root spread of PB1 and Dular under +Pi and −Pi condition. Lower panel shows points of root emergence from the side and bottom of basket.

(G) Percentage of roots emerging from side (outer) and bottom (inner) regions of basket under +Pi and (H) −Pi. Scale bar = 5 cm. **p < 0.01; ***p < 0.001; ns not

significant as determined by student’s t-test.

Frontiers in Plant Science | www.frontiersin.org 4 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

TABLE 1 | Plant growth and Pi accumulation in Dular and PB1.

Genotype Growth condition Root length (cm) Shoot length

(cm)

Shoot surface

Area (cm2)

Root Pi

nmolPi/mg FW

Shoot Pi

nmolPi/mg FW

Dular +Pi 7.03 ± 0.09 24.25 ± 0.39 5.14 ± 0.04 1.65 ± 0.07 9.28 ± 0.37

−Pi 9.38 ± 0.17 21.9 ± 0.33 4.95 ± 0.03 0.15± 0.01 0.90 ± 0.06

PB1 +Pi 8.31 ± 0.18 21.93 ± 0.30 6.83 ± 0.33 2.27 ± 0.13 11.98 ± 0.41

−Pi 7.41 ± 0.15 15.36 ± 0.33 3.17 ± 0.10 0.03 ± 0.002 0.73 ± 0.05

Effects Root length Shoot length Shoot surface

area

Root Pi Shoot Pi

Genotype (G) 0.15ns 1.7E-15# 0.8ns 2.2E-05# 0.001***

Treatment (T) 0.01** 9.6E-20# 0.001*** 9.7E-49# 5.1E-39#

G × T 4.8E-12# 1.3E-05# 0.06ns 2.2E-06# 2.5E-06#

“+” and “−” indicate 320µM Pi and 1µM Pi in nutrient medium, respectively. Values represent mean of at least 20 replicates.**; ***; “#” indicates highly significant; ns indicate not

significant.

Gene Expression Profiles Under PiDeficiencyTo understand the molecular basis of differential low Pi tolerancein two genotypes, we performed global transcriptome analysisusing microarrays. Three-way ANOVA analysis on microarraydata distinguished the low Pi responsive genes from thoseaffected by genotype and tissues independent of Pi levels. Outof total 57,381 probe sets on GeneChip, an overall 39,243 probesets “significantly expressed’ at p ≤ 0.05 in either of the tissues,genotypes and Pi treatment. These 39,243 probe sets represented16,458 unique genes. Of these, expressions of 12,619, 8135, and13,541 genes were influenced by the main effects of genotype,treatment and tissue, respectively. Whereas, expressions of 4230,8466, 3379, and 1181 genes were influenced by the effect ofinteractions of Genotype × Phosphate (G × Pi), Genotype ×

Tissue (G × T), Tissue × Phosphate (T × Pi) and Genotype ×Tissue × Phosphate (G × T × Pi), respectively (SupplementaryFigure 1). Out of 8135 low Pi affected genes (SupplementaryTable 3), 77% and 84% were also influenced by the variabilityof genotype and tissue, respectively. Further, 55% of genes wereinfluenced by the two way interaction of G × T while 32% and33% of Pi responsive genes were influenced by G × Pi andT × Pi interactions, respectively. Only 11% of Pi responsivegenes were influenced by the 3-way interactions of G × T × Pi.This analysis revealed high order complexity and intricacy of allthree variables. Out of 8135 genes, 1457 genes were up-regulated(−Pi/+Pi) and 1013 were down-regulated (−Pi/+Pi) in Dularroots whereas, in PB1 roots, 893 and 710 genes were significantly(≥2-fold) up and down-regulated, respectively. Shoot tissues, onthe contrary, showed a relatively lower number of genes withaltered expression (Figure 2A). This indicates more dynamicgene expression in root tissues under Pi deficiency, especiallyin the low Pi tolerant Dular genotype. Out of 12,619 genesinfluenced by the genotype factor (Dular/PB1), 4708 and 3908genes were differentially regulated in root and shoot tissuesunder Pi deficiency, respectively. Comprehensive analysis ofdifferentially expressed genes showed perturbations in a variety ofcellular responses directed toward mitigation of Pi deficiency and

other regular plant processes (Supplementary Figure 2). Further,we found many Phosphate Starvation Response (PSR) geneslike PHO1, phosphatase, SPX domain containing sequences andothers among the differentially expressed genes in both genotypes(Supplementary Figure 3).

Results of the microarray experiments were also successfullyvalidated using qRT-PCR (Figure 2B; Supplementary Table 4) forrandomly selected genes with a correlation coefficient >0.94.

Expression Profile of Known PSR Genes inDular and PB1 Under Pi DeficiencyMany PSR genes have been identified and characterized fortheir roles in improving plant Pi deficiency tolerance. Wecompiled a list of 40 such common genes by screening thepublished literature and analyzed their expression behaviorin our data (Figure 3). Interestingly, >30 of them wereprominently up-regulated in roots, with higher expression levelsin Dular than PB1. OsIPS1 and a ser/thr phosphatase wereup-regulated >900-fold in Dular roots. Other Pi responsivegenes like SPX domain containing proteins, PHO genes,acid phosphatases, phosphoethanolamine phosphatase, andnucleotide pyro-phosphatase were also highly up-regulated inDular roots as compared to PB1 roots. However, Pi transporterand purple acid phosphatase (Os08g17784) showed greater up-regulation in PB1 roots. We also compiled publically availablegene expression data of different rice genotypes under Pideficiency. Comparison of our data with this list revealed anumber of common and many unique PSR genes identified inour study (Supplementary Tables 5, 6).

Expression Analysis of Genes Involved inInternal Pi UtilizationInduction of Transcripts Involved in Glycolytic

BypassesUnder Pi deprivation, plant cells bypass ATP or Pi dependentbiochemical reactions of sugar metabolism using alternate Pi-independent enzymes. Genes involved in three such glycolyticbypasses were altered in both genotypes with higher induction

Frontiers in Plant Science | www.frontiersin.org 5 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

FIGURE 2 | Gene expression profiles of 15-days-old PB1 and Dular seedlings under Pi deficiency. (A) Venn diagram showing unique and commonly

regulated low Pi responsive genes (FC ≥ 2) in PB1 and Dular. Numbers outside and inside the parenthesis indicate up and down-regulated genes, respectively. (B)

Confirmation of microarray results with qPCR experiments. r-value represents the correlation coefficient between log2transformed fold changes of microarray and

qPCR results.

in Dular (Figure 4). The first bypass event is catalyzed byPPi-dependent phosphofructokinase (PPi-PFK), which convertsthe fructose-6 phosphate to fructose 1, 6-bisphosphate withoutconsumption of ATP. Two PPi-PFK genes, Os02g48360 andOs06g22060, were up-regulated only in Dular roots. However,one such gene (Os08g25720) was also induced in the shootof PB1. In the second bypass, NADP-GAPDH circumventsthe Pi requiring NAD-GAPDH enzyme for the formation of1, 3-bisphosphoglycerate. The gene (Os12g12590) encodingNADP dependent aldehyde dehydrogenase was specifically up-regulated in Dular roots. Further downstream in glycolysis, onemore alternate pathway operates through phosphoenol pyruvatecarboxylase (PEPC) and malate dehydrogenase (MDH) to bypassthe Pi utilizing pyruvate kinase. In both Dular and PB1, aPEPC encoding gene, Os09g14670, and one MDH encodinggene, Os08g33720 were prominently induced under low Pitreatment. Such adjustment mechanisms were induced in boththe genotypes; with a biased overrepresentation in Dular.

Dular Showed Higher Expression of Lipid

Remodeling GenesPi deficiency leads to global membrane lipid remodeling inorder to release Pi from membrane phospholipids (Table 2).This involves two major steps, degradation of phospholipidsto diacylglycerol and subsequent conversion into galactolipidsand sulfolipids. Phospholipid hydrolysis is either mediatedby phospholipase C (PLC) in a single step reaction or intwo-steps by phospholipase D (PLD) and phosphatidatephosphatase (PAH) (reviewed in Nakamura, 2013). PLCand PLD encoding genes were significantly up-regulated inDular and PB1 roots and PB1 shoot under low Pi (Table 2).Further, two PAH genes (Os11g40080, Os05g38720) werealso preferentially up-regulated in Dular, especially in roottissue. In a second step, the major product of phospholipidhydrolysis, diacylglycerol is channeled into biosynthesis ofgalactolipids, monogalactosyldiacylglycerol (MGDG), anddigalactosyldiacylglycerol (DGDG) in plastid membrane

(Nakamura, 2013). We found up-regulation of MGDG synthase(Os08g20420) in both genotypes with relatively higher expressionin Dular roots. Noticeably, higher up-regulation in Dular shootswas also observed for genes encoding DGDG (Os03g11560,Os04g34000).

Genes involved in sulfolipid biosynthesis, UDP-sulfoquinovose synthase (SQD1) and sulfolipid synthase (SQD2),also showed significant induction in Dular and PB1, especially inroots. SQD2 encoding gene, Os01g04920, was expressed at ≥80-fold in Dular roots and 56-fold in PB1 roots under Pi deficiency.In an alternative pathway, phospholipids hydrolysis is mediatedby glycerophosphodiester phosphodiesterases (GDPD). Furtheranalysis revealed four genes encoding GDPD were up-regulated.Of these, Os03g40670 was highly up-regulated in Dular roots(Table 2). An overall fairly high induction of lipid metabolismgenes in Dular indicates better cellular Pi homeostasis in thisgenotype under Pi starvation. To confirm this hypothesis,we analyzed the effect of “Genotype” (G) factor by drawing acomparison of absolute level of transcripts (signal intensities)in Dular with respect to PB1 (i.e., Dular/PB1) under bothPi deficient and sufficient conditions (Table 2). Our analysisconfirmed that Dular has higher absolute expression of lipidremodeling genes in shoot and to a lesser extent in roots aswell, under Pi deficient conditions when compared to PB1. It isnoteworthy from Table 2 that under the sufficient Pi conditionDular has lesser transcripts than PB1 in both root and shoot.Our analysis revealed that expression of lipid remodelinggenes more strongly regulated by the genotype than Pi stress.Thus, these genes can contribute to the low Pi tolerance ofDular.

Lipid Phenotyping Revealed Greater Metabolic

Flexibility of Dular Under low PiWe validated the alterations in transcript abundance of lipidremodeling genes by comprehensive metabolic phenotyping(Supplementary Table 7). Our analysis of relative lipidcomposition (+Pi/−Pi) revealed decreases in phospholipids such

Frontiers in Plant Science | www.frontiersin.org 6 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

FIGURE 3 | Expression profile of known PSR genes under Pi deficiency. Gene list was compiled from published reports. 1, Dular root; 2, PB1 root; 3, Dular

shoot; 4, PB1 shoot.

as PC (Phosphatidylcholine), PE (Phosphatidylethanolamine),and PI (Phosphatidylinositol), with a subsequent increase ingalactolipids (e.g MGDG, DGDG) under low Pi (Tables 3, 4).Our lipidome analysis also captured the genotypic variability

(Dular/PB1) in levels of phospholipids. Under Pi sufficientand deficient conditions, Dular shoots showed higher contentsof phospholipids as compared to PB1 shoots (Table 3). Onthe other hand, PB1 roots and shoots showed a significant

Frontiers in Plant Science | www.frontiersin.org 7 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

FIGURE 4 | Differential expression of key genes involved in carbohydrate metabolism under Pi deficiency. Expression profiles of differentially expressed

genes regulating key metabolic reactions are shown adjacent to their enzyme products. 1, Dular root; 2, PB1 root; 3, Dular shoot; 4, PB1 shoot.

decrease in phospholipid content as compared to Dular underPi deficiency. This indicates faster degradation of phospholipids(i.e., degradation of major proportion of phospholipids) in PB1under Pi deficiency from its limited reserves. This further impliesthat Dular has a greater phospholipid reserve to release Pi underlow Pi and maintains membrane integrity at the same time.Moreover, there was significant accumulation of phospholipidsPC and PE in Dular roots under low Pi. All these evidences

reveal greater potential of Dular to cope with low Pi stress thanPB1.

Interestingly, low Pi-tolerant Dular also showed increasedaccumulation of galactolipids, DGDG in both roots and shootsunder low Pi as compared to PB1.While, increased accumulationof MGDG was found in Dular shoots under low Pi (−Pi/+Pi)(Table 4). Genotypic analysis (Dular/PB1) revealed increasedaccumulation of MGDG and DGDG in Dular shoots as

Frontiers in Plant Science | www.frontiersin.org 8 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

TABLE 2 | Expression profile of differentially expressed genes encoding key enzymes in lipid metabolic pathway.

Gene ID Functional annotation A. fold change (−Pi/+Pi) B. fold change (Dular/PB1)

Dular root PB1 root Dular shoot PB1 shoot Root +Pi Root −Pi Shoot +Pi Shoot −Pi

Os01g19390 ATS 2.94 1.45 1.08 1.02 −1.19 1.70 1.01 −1.04

Os01g22560 ATS −1.06 −1.01 −7.79 1.56 1.07 1.03 −7.98 135.54

Os05g38350 ATS 5.49 1.77 1.09 1.67 −7.65 −2.47 −3.76 1.25

Os05g38720 PAH 4.37 1.56 1.50 1.36 −1.73 1.61 −1.69 1.16

Os11g40080 PAH 4.01 2.87 3.12 2.20 −1.13 1.23 1.15 1.47

Os03g30130 PLC 4.13 2.63 1.81 2.93 1.82 2.86 1.70 1.44

Os12g37560 PLC 1.56 2.00 1.15 1.11 −1.72 −2.22 −1.26 −1.44

Os06g40170 PLD 3.14 4.02 1.39 2.08 −1.46 −1.87 −2.60 2.43

Os06g40180 PLD 3.47 2.73 1.30 1.37 −1.24 1.03 −5.63 −2.64

Os08g20420 MGD2 627.68 190.77 110.82 123.37 −1.05 3.13 −1.04 16.43

Os03g11560 DGD2 4.20 2.36 3.46 2.03 −1.25 1.43 −1.49 1.35

Os04g34000 DGD2 16.95 8.70 23.15 14.54 −2.80 −1.44 −3.24 5.67

Os05g32140 SQD1 3.51 8.23 2.45 2.89 2.02 −1.16 1.17 2.28

Os01g04920 SQD2 82.66 56.18 16.50 10.19 1.20 1.76 −1.07 3.38

Os07g01030 SQD2 4.40 5.31 2.48 1.81 1.37 1.14 −1.45 2.40

Os01g55780 GDPD 2.31 1.53 1.67 1.28 −1.13 1.33 −1.15 1.08

Os02g31030 GDPD 4.97 4.38 34.50 60.46 1.15 1.30 1.61 1.73

Os03g40670 GDPD 403.33 80.96 2.89 7.23 −6.06 −1.22 1.46 3.06

Os08g42390 GDPD 3.54 2.96 14.35 27.58 2.19 2.62 1.38 1.46

ATS, Acetyl Transferase; PAH, phosphatidate phosphatase; PLC, Phospholipase C; PLD, Phospholipase D; SQD1/2, UDP-sulfoquinovose synthase1/2; GDPD, glycerophosphodiester

phosphodiesterases.

compared to PB1 under Pi deficiency (Table 4). Noticeably,these observations are in high concordance with transcriptomeresults of shoot tissue in both the genotypes. However, PB1root tissue had a higher accumulation of galactolipids thanDular under Pi deficiency (Table 4). To assess the impact of thisobservation, we further analyzed the lipid data in terms of thirdfactor, “Tissue” (Shoot/Root; Table 4). Interestingly, we foundhigher accumulation of galactolipids in shoots than roots. Thisdifferential accumulation accounts to the fact that photosyntheticleaves contain chloroplasts with larger surface area as comparedto roots with rudimentary plastids. Galactolipids form the bulkof these shoot thylakoid membranes as they play important rolein light reactions (Dormann, 2007). Thus, higher accumulationof galactolipids in Dular shoots might outpace the advantage ofhaving higher accumulation of galactolipids in PB1 roots underPi deficiency.

Expression Analysis of Genes Involved inPi AcquisitionInduction of RSA, Cell Wall Loosening, and

Biosynthesis GenesPi deficiency generally alters root architecture for increasingsoil foraging capacity. Therefore, we compared our datawith available transcriptome data of rice roots in differentdevelopmental stages and tissues (Takehisa et al., 2012). Ouranalysis revealed upregulation of 64 genes involved in lateralroot development in Dular roots, but upregulation of only33 genes in PB1 (Supplementary Table 8). Furthermore, cellwall loosening and biosynthesis are essential processes for

modulating RSA. A comparison of our data with the CellWall Navigator Database (Girke et al., 2004) revealed highernumbers of up-regulated root growth and cell wall biosynthesisgenes in Dular roots (66) as compared to PB1 roots (34).These genes include expansins, glycosyl transferases, glycosylhydrolases and cellulose synthase (Supplementary Table 9 andSupplementary Figure 4). A comparative analysis with rice GTgenes database (Cao et al., 2008) showed 109 differentiallyexpressed genes in both the genotypes (Supplementary Table10). Interestingly, Dular roots showed a higher number ofup-regulated GT genes (82) than PB1 roots (37). Five ofthe xyloglucan endotransglucosylases (Os04g51460, Os06g48200,Os07g34580, Os03g02610, and Os03g63760) were also foundsignificantly upregulated in Dular roots under Pi deficiency. Thisagain showed the general tendency of higher up-regulation oftranscripts related to root growth and cell wall loosening in Dularroot as compared to PB1 roots.

Increased Expression of Acid Phosphatases and

Organic Acids Genes in DularPhosphatases, ribonucleases and organic acids release the Pifrom organic/inorganic compounds in the rhizosphere underits deficiency. We found 266 various types of phosphatasesand hydrolases differentially expressed in Dular and PB1,out of which, 128 and 95 genes were up-regulated inDular and PB1 roots, respectively (Supplementary Table 11).Interestingly, up-regulation of ser/thr phosphatase (Os10g02750)and phosphocholine phosphatase (Os01g52230) genes was ashigh as 909-and 552-fold in Dular roots in comparison to

Frontiers in Plant Science | www.frontiersin.org 9 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

TABLE 3 | LC-QTOF analysis of relative concentrations of phospholipids under Pi deficiency.

m/z tR Annotation A. peak ratios (−Pi/+Pi) B. peak ratios (Dular/PB1)

Dular

root

PB1

root

Dular

shoot

PB1

shoot

Root

+Pi

Root

−Pi

Shoot

+Pi

Shoot

−Pi

734.57 5.44 PC (16:0/16:0) [M+H]+ 0.08 0.04 0.1 0.05 1.12 2.15 2.29 4.26

732.55_754.54 4.92_4.79 PC (32:1) [M+H]+_PC

(32:1) [M+Na]+

0.39 0.06 0.16 0.13 0.41 2.84 2.24 2.83

752.52_730.54 4.43_4.44 PC (32:2) [M+Na]+_PC

(32:2) [M+H]+

0.27 0.09 0.12 0.04 0.62 1.81 2.88 10.13

744.55_766.54 4.73_4.77 PC (33:2) [M+H]+_PC

(33:2) [M+Na]+

0.44 0.07 0.23 0.06 0.44 2.71 1.54 6.18

742.54 4.34 PC (33:3) 0.5 0.07 0.24 0.06 0.36 2.49 2.95 11.68

760.58 5.76 PC (34:1) A [M+H]+ 0.32 0.14 0.08 0.32 0.5 1.15 2.99 0.79

760.58_782.57 5.52_5.52 PC (34:1) B [M+H]+_PC

(34:1) B [M+Na]+

0.3 0.1 0.17 0.05 0.46 1.38 2.66 9.08

758.57 5.19 PC (34:2) A [M+H]+ 0.3 0.07 0.22 0.08 0.68 2.71 2.9 7.52

780.55_758.57 5.03_5.03 PC (34:2) B [M+Na]+_PC

(34:2) B [M+H]+

0.41 0.18 0.32 0.09 0.72 1.69 1.82 6.72

756.56 4.95 PC (34:3) A [M+H]+ 0.47 0.41 0.35 0.25 0.25 0.28 3.15 4.53

756.55_778.54 4.61_4.61 PC (34:3) B [M+H]+_PC

(34:3) B [M+Na]+

0.45 0.17 0.45 0.1 0.61 1.64 2.77 13.03

754.54 4.26 PC (34:4) 0.42 0.1 0.15 0.15 0.54 2.34 5.37 5.29

772.58 5.05 PC (35:2) A [M+H]+ 0.48 2.13 1.8 0.39 2.33 0.52 0.6 2.81

772.59 5.39 PC (35:2) B [M+H]+ 0.31 0.32 0.22 0.3 0.71 0.7 2.42 1.78

770.57 4.90 PC (35:3) B [M+H]+ 0.39 0.12 0.31 0.52 0.66 2.09 4.26 2.55

788.62_810.60 6.17_6.17 PC (36:1) [M+H]+_PC

(38:4) B [M+H]+

0.14 0.15 0.27 0.1 1.02 0.98 1.72 4.53

786.60_808.58 5.63_5.64 PC (36:2) [M+H]+_PC

(36:2) [M+Na]+

0.18 0.14 0.32 0.09 0.59 0.72 1.74 6.15

784.59 5.13 PC (36:3) A [M+H]+ 0.5 0.15 0.17 0.05 0.42 1.45 2.44 7.83

804.55_782.57 4.65_4.64 PC (36:4) A [M+H]+_PC A

(36:4) [M+Na]+

0.52 0.19 0.33 0.11 0.57 1.59 1.6 5

804.55 5.04 PC (36:4) B [M+Na]+ 0.21 0.41 0.83 0.47 1.38 0.69 1.07 1.92

780.55 4.29 PC (36:5) [M+H]+ 0.46 0.11 0.39 0.12 0.37 1.49 2.61 8.76

778.54 3.91 PC (36:6) [M+H]+ A 0.39 0.08 0.35 0.12 0.23 1.14 6.78 19.93

800.62 6.04 PC (37:2) [M+H]+ 0.11 0.21 0.67 0.88 1.25 0.66 1.92 1.46

777.55_794.58 4.73_4.72 PC (37:5) [M+H]+_PC

(37:5) [M+NH4]+

0.2 0.14 1.61 0.61 0.43 0.6 0.62 1.63

808.58_808.58 4.78_5.03 PC (38:5) A [M+H]+_PC

(38:5) A [M+H]+

0.44 0.49 0.23 0.25 1.03 0.91 2.22 2.05

806.57 4.69 PC (38:6) A [M+H]+ 0.57 0.38 0.34 0.22 0.91 1.36 1.68 2.58

806.57 5.10 PC (38:6) B [M+H]+ 0.52 0.16 0.15 0.09 0.42 1.37 2.86 5.02

742.58 5.15 PC (p-34:2) or PC (o-34:3)

[M+H]+

0.7 0.99 0.44 0.67 1.34 0.95 1.46 0.96

798.64 6.44 PC (p-38:2) or PC (o-38:3)

[M+H]+

0.48 0.63 0.52 0.79 1.02 0.77 1.35 0.89

794.60 5.78 PC (p-38:4) B or PC

(o-38:5) B [M+H]+

0.34 0.26 0.45 0.7 0.63 0.82 1.73 1.13

716.52_738.51 5.18_5.19 PE (34:2) [M+H]+_PE (34:2)

[M+Na]+

0.14 0.11 0.08 0.03 0.89 1.18 2.68 8.04

740.53 4.80 PE (36:4) [M+H]+ 0.21 0.1 0.16 0.05 0.48 1 2.36 7.13

714.51 5.21 PE (34:2) [M−H]− B 0.16 0.12 0.1 0.03 0.83 1.11 2.6 7.87

740.52 5.29 PE (36:3) [M−H]− B 0.4 0.21 0.5 0.35 0.59 1.1 2.49 3.56

818.59 6.11 PE (38:1) [M+HCOO]− A 0.76 1.65 0.49 0.48 0.91 0.42 1.39 1.45

747.52 5.04 PG 34:1 [M−H]− 0.04 0.03 0.23 0.15 0.52 0.57 1.94 2.97

(Continued)

Frontiers in Plant Science | www.frontiersin.org 10 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

TABLE 3 | Continued

m/z tR Annotation A. peak ratios (−Pi/+Pi) B. peak ratios (Dular/PB1)

Dular

root

PB1

root

Dular

shoot

PB1

shoot

Root

+Pi

Root

−Pi

Shoot

+Pi

Shoot

−Pi

745.50 4.62 PG 34:2 [M−H]− 0.11 0.06 0.07 0.13 0.53 0.87 2.58 1.41

819.53 4.32 PG 40:7 [M−H]− 0.14 0.13 1.01 0.51 0.46 0.48 0.98 1.94

835.53 4.92 PI (34:1) [M−H]− 0.31 0.21 0.46 0.44 0.66 0.97 0.88 0.93

901.51 and

833.52

4.48 and

4.45

PI (34:2) [M+NaHCO2]− &

PI (34:2) [M−H]−

0.21 0.23 0.15 0.13 0.77 0.71 0.96 1.12

857.52 4.14 PI 36:4 [M−H]− 0.18 0.28 0.19 0.26 0.83 0.52 0.64 0.47

PC, PhosPC: photidylcholine; PE, Phosphotidylehanolamine; PG, Phosphotidylglycerol; PI, Phosphotidylinositol.

TABLE 4 | LC-QTOF analysis of relative concentrations of galactolipids under Pi deficiency.

m/z tR Annotation A. peak ratios (-Pi/+Pi) B. peak ratios (Dular/PB1) C. peak ratios (Shoot/Root)

Dular

Root

PB1

Root

Dular

Shoot

PB1

Shoot

Root

+Pi

Root

-Pi

Shoot

+Pi

Shoot

-Pi

Dular

+Pi

PB1

+Pi

Dular

-Pi

PB1

-Pi

941.6176_

936.6609

5.538_5.51 DGDG 34:1 [M+Na]+

DGDG 34:1 [M+NH4]+

1.54 1.71 1.19 1.32 0.60 0.54 0.51 0.46 0.94 1.10 0.73 0.85

962.6771 5.69 DGDG 36:2 [M+NH4]+ 2.08 1.81 3.01 2.25 0.70 0.80 0.51 0.69 2.21 3.02 3.20 3.74

965.6163_

960.6603

5.206_5.18 DGDG 36:3 [M+Na]+

DGDG 36:3 [M+NH4]+

3.26 2.40 1.96 1.02 0.67 0.91 0.61 1.18 6.45 7.09 3.88 3.01

961.5859_

956.6313

4.267_4.27 DGDG 36:5 [M+Na]+

DGDG 36:5 [M+NH4]+

0.90 0.56 2.26 2.08 0.41 0.65 0.63 0.69 5.42 3.46 13.68 12.82

954.6155 3.92 DGDG 36:6 [M+NH4]+ 0.16 0.15 1.17 0.84 0.36 0.38 0.91 1.26 28.33 11.33 209.7 63.02

779.5652 6.03 MGDG 34:1 [M+Na]+ 0.68 2.63 1.01 0.37 0.78 0.20 0.61 1.69 0.86 1.10 1.29 0.15

777.5495 5.54 MGDG 34:2 [M+Na]+ 0.61 2.68 1.73 0.74 1.25 0.29 0.62 1.47 0.49 0.99 1.39 0.27

773.5189 4.59 MGDG 34:4 [M+Na]+ 0.08 0.03 0.96 0.16 0.44 1.01 0.44 2.57 1.29 1.30 16.20 6.38

803.5664_

798.6027

5.681_5.54 MGDG 36:3 [M+Na]+

MGDG 36:3

[M+NH4]+

0.78 1.02 1.43 0.37 0.39 0.30 0.70 2.71 3.00 1.69 5.47 0.61

799.5335 4.70 MGDG 36:5 [M+Na]+ 0.28 0.33 1.37 0.74 0.62 0.53 0.70 1.30 2.27 2.02 10.91 4.49

797.5183_

792.5621

4.301_4.31 MGDG 36:6 [M+Na]+

MGDG 36:6

[M+NH4]+

0.10 0.11 1.02 0.47 0.48 0.44 0.96 2.11 6.76 3.34 71.60 14.87

770.5749 5.11 MGDG 34:3

[M+NH4]+

0.38 0.47 1.75 0.39 0.54 0.43 0.61 2.75 0.94 1.10 0.73 0.85

MGDG, Monogalactosyldiacylglycerol; DGDG, Digalactosyldiacylglycerol.

318-and 47-fold in PB1 roots, respectively. In-gel APase assayalso revealed the higher activity of APases under Pi deficiency inDular roots as compared to PB1 (Figure 5A). An earlier reportedPi inducible E1 isoform of APase, OsPAP10, was found to bestained with equal intensity under Pi deficiency. The gene for thesame isoform also showed almost equal upregulation of 9.16 and9.62 in Dular and PB1 roots, respectively.

We also found prominent upregulation of a PEPC encodinggene (Os09g14670) in Dular roots and shoots (42-and 11-fold) as compared to PB1 roots and shoots (2.7-and 4.2-fold)which catalyze the synthesis of oxaloacetate and release Pias a by-product (Supplementary Table 11 and Figure 5B).Additionally, genes involved in malic acid metabolism,(MDH Os04g46560, Os08g33720) were also up-regulatedin roots of both genotypes. Moreover, genes for ribonucleases

(RNases) (Os08g33710, Os01g67180, Os01g67190) and acitrate transporter (Os10g31040) were also up-regulatedmore prominently in Dular roots. This comparative studyof phosphatases, RNases, and organic acids encoding genesrevealed a potential ability to solubilize soil-bound Pi byDular under Pi deficiency (Supplementary Table 11 andFigure 5).

Higher Induction of Pi Transporter Genes in PB1We observed that 104 transporter encoding genes wereupregulated in Dular roots while only 52 were in PB1 underPi stress (Supplementary Table 12). Out of 26 potential Pitransporters in rice, 5 Pi transporters, including the highaffinity Pi transporter OsPT6, were induced (SupplementaryFigure 5). Interestingly, high affinity Pi transporters like OsPT6

Frontiers in Plant Science | www.frontiersin.org 11 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

FIGURE 5 | Effect of Pi deficiency on APases and genes involved in organic acids metabolism. (A) In-gel APase activity of proteins extracted from PB1 and

Dular roots under +Pi and -Pi conditions. (B) Expression profile of APases showing higher expression of APases in Dular and PB1 under +Pi and -Pi. (C) Differential

expression of genes involved in organic acid metabolism in Dular and PB1 under Pi deficiency. 1, Dular root; 2, PB1 root; 3, Dular shoot; 4, PB1 shoot.

(Os08g45000) showed higher upregulation in PB1 roots ascompared to Dular. Analysis of absolute transcript expression(Dular/PB1) revealed a higher expression (3.6) of OsPT6in PB1 roots under Pi deficiency as compared to Dular.Notably, our data also reflect the transcript induction of theOsPHO1-3 (Os06g29790) gene involved in Pi loading intoxylem. Transcripts of four SPX domain containing proteins(Os03g29250, Os10g25310, Os06g40120, and Os02g10780) werealso up-regulated in both Dular and PB1 roots.

Identification of SNPs and Indels BetweenDular and PB1 in Differentially ExpressedPSR GenesWe also identified SNPs and InDels presents between Dularand PB1 in 8135 differentially regulated genes under low Piconditions using available genome sequence (Mehra et al., 2015;Supplementary Table 13). 75244 genic SNPs and 16400 genicInDels were identified between Dular and PB1; however, 17739coding SNPs and 1123 coding InDels were discovered. 5′ UTRsand 3′ UTRs of Dular and PB1 were also analyzed for SNPs andInDels which yielded 3978, 9048 SNPs and 1772 and 2455 InDels,respectively. Functional classification of these SNPs and InDelsrevealed 10 startloss and 18 stoploss SNPs between Dular andPB1. Moreover, 69 non-sense and 7337 missense as well as 7805silent SNPs were identified. Further, 217 large effect frameshiftInDels between Dular and PB1 were also observed. Additionally,we also analyzed the SNPs and InDels present in 2kb upstreampromoter regions of key PSR genes (Supplementary Table14). Interestingly, numerous SNPs and InDels were found inpromoter regions of lipid remobilizing genes (SQD2, GDPD, ATS,PAH, MGD2, DGD2, PLD), a high affinity phosphate transporter(OsPT6), SPX domain containing proteins, various organic acidgenes (lactate/malate dehydrogenase, PEPC) and purple acidphosphatases between Dular and PB1 (Supplementary Table 14).

DISCUSSION

We investigated the potential low Pi tolerance mechanismsusing comparative morphophysiological, transcriptomics, andlipidomics approaches in a low Pi sensitive modern rice genotype(PB1) and a tolerant genotype, Dular. Pusa Basmati-1 (PB1) is thefirst high yielding semi-dwarf Basmati variety which yields∼50–52 q/ha and shows improved growth parameters with highnutrient supply (Sharma et al., 2012). On the other hand, Dularis a traditional genotype with very low yield potential of about7–22 q/ha and has been shown to be one of the most tolerantrice genotypes for Pi deficiency in field conditions (Wissuwa andAe, 2001; Chin et al., 2010). Low Pi tolerance can be achievedthrough better Pi acquisition and efficient cellular Pi utilization(Wang et al., 2010; Rose et al., 2011). Higher “Pi acquisitionefficiency” corroborates to enhanced Pi uptake, root architecturalmodifications and solubilisation of bound Pi in soil whereas“Pi utilization efficiency” refers to efficient remobilization ofcellular Pi. Generally, modern genotypes are believed to be highly“Pi responsive” due to higher Pi uptake, whereas, traditionalgenotypes like Dular are “Pi efficient” (Gerloff, 1977; Wissuwaand Ae, 2001). However, insufficient molecular evidences exist tosupport this notion. Therefore, we employed transcriptome andlipidome profiling of both contrasting genotypes to reveal the lowPi tolerance mechanisms of a Pi efficient genotype.

Three way ANOVA analysis of microarray data revealedthat most of the transcripts were influenced by tissue andgenotype. This suggests that most of the Pi responsive genes wereexpressed at different background levels in selected genotypeand tissue types. We found two- and five-fold higher numbersof differentially expressed transcripts affected by the genotypefactor as compared to the Pi treatment factor in Dular rootand shoot tissues, respectively. Thus, there exists a significantgenetic variability between the selected genotypes which providesa novel aspect to unravel the low Pi tolerance mechanisms. Our

Frontiers in Plant Science | www.frontiersin.org 12 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

integrated transcriptome, lipidome, and morphological analysisrevealed that Dular showed multifaceted tolerance mechanismswhich involve RSA modulation and lipid remobilization.However, earlier studies conducted by Pariasca-Tanaka et al.(2009) had shown RSA modulation as the only mechanism forlow Pi tolerance of NIL6-4 (introgressed with Pup1 QTL fromlow Pi tolerant aus genotype, Kasalath). This was probably dueto less genetic variability between NIL6-4 and low Pi sensitiverecurrent parent, Nipponbare. Intriguingly, both PB1 and Dularcarry the Pup1 candidate gene Pstol1 (data not shown). Therefore,present study becomes more rational to look into the otherlow Pi tolerance mechanisms. In another study, Oono et al.(2013) found that low Pi induced genes are more up-regulatedin tolerant aus genotypes, Kasalath than sensitive genotypes.Our study also confirms the same hypothesis in another ausgenotype Dular and proved the importance of transcriptionalregulation for low Pi adaptation. Furthermore, similar to earlierobservation in Kasalath (Oono et al., 2013), Dular also showedreduced inorganic P content which reflects on potential efficientP utilization.

RSA modulations like increased root hairs, lateral rootsand higher turnover of roots are important traits involved inadaptation to Pi deficiency (Lynch, 2011; Gamuyao et al., 2012;Pandey et al., 2013). Sensing of Pi deficiency at the root tipactivates the downstream Pi responsive genes associated withRSA modulation (Svistoonoff et al., 2007). Our transcriptomedata also showed higher induction of RSA genes (glycosylases,expansins, XTHs) in Dular roots in comparison to PB1 underPi deficiency. Phenotypic analysis of Dular and PB1 genotypesalso depicted apparent differences in the root architecture ofboth genotypes. In an earlier study, we found increased lateralroots and root hair growth in Dular under low Pi as comparedto PB1 (Figure 1C, Mehra et al., 2015). Both traits bestowenhanced Pi acquisition by increasing the absorptive surfacearea under Pi limiting condition. Additionally, Dular rootsshowed a significant increase in root length under low Pi. Thisquantitative trait has been linked with Pi deficiency tolerance insome rice genotypes (Shimizu et al., 2004). Increased root lengthis associated with longer and more branched roots per unit ofroot dry matter (Hill et al., 2006). Root elongation enhances theporosity and oxygen release capacity of plants which leads toiron oxidation and release of protons. This causes an increasein rhizosphere acidity that helps in solubilisation of soil Picompounds (Kirk and Du, 1997). Therefore, increase in rootlength helps in surviving in Pi poor soil. Our analysis of rootspread showed that PB1 possesses shallow root system whichsuggests that it has been selected on nutrient rich soils for higherPi acquisition attributing to its higher yield. Since, PB1 is adrought sensitive genotype (Mutum et al., 2013) shallow rootsystem may contribute to its high drought sensitivity. On theother hand, we observed a steeper-angled root system in Dularwhich also likely contributes to its drought avoidance mechanismas Dular is a traditional upland genotype (Henry et al., 2011).Moreover, upland systems are co-limited for water and Pitogether. Therefore, Dular has been co-selected for both thetraits i.e., low Pi tolerance and drought avoidance. Therefore, thistraditional “Pi-efficient” genotype can also be further exploited

to target the dual problem of drought and low Pi tolerance. Wedidn’t find significant change in root spread for either genotypecaused by low Pi. This indicates that root spread behavior mightbe governed by genotype rather than low Pi stress. Interestingly,we also observed greater cortical aerenchyma formation inDular than PB1 root under Pi deficiency (Supplementary Figure6). This adaptive trait allows reallocation of resources towardformation of new roots with minimum additional metabolic cost(Lynch, 2011; Postma and Lynch, 2013). Noticeably, under Pideficiency, Dular roots also showed higher induction of XTHs(Xyloglucan endotransglucosylase) genes known to be involvedin aerenchyma and root hair formation (Vissenberg et al., 2001).It is noteworthy that NIL6-4 with introgressed Pup1 locus didn’tshow such advantage in root hair and aerenchyma formationdespite induction of XTHs. This implies that these complex traitsinvolve several molecular regulators influenced by environmentand genotype. The integration of these root traits may lead tosynergistic interactions that substantially increase P uptake (Yorket al., 2013).

Metabolic plasticity plays an important role in conservingcellular Pi (Plaxton and Tran, 2011) and thereby enhances Piutilization efficiency. In our data, many key genes involved inglycolytic bypasses were significantly up-regulated in Dular thusefficiently mitigating the demand of Pi under its deficiency.Cellular Pi pools can also be conserved by replacing membranephospholipids with galactolipids and sulfolipids (Nakamura,2013; Okazaki et al., 2013). A set of membrane lipid remodelinggenes, including like PLC, PLD, and GDPD, were found tobe induced in Dular at relatively higher levels than PB1under low Pi. Further, low Pi induction of lipid remodelinggenes and accumulation of corresponding metabolic lipid bio-markers (MGDG and DGDG) indicate that lipid remobilizationis also one of adaptive strategies for low Pi tolerance inthe traditional genotype. Analysis of phospholipid compositionshowed higher degradation of phospholipids in PB1 underlow Pi. However, the total amount of phospholipids degradedis more in Dular (analogy model, Supplementary Figure 7).Genotypic comparison of Dular and PB1 revealed a greater poolof phospholipids in Dular shoots under both low and high Pias compared to PB1. This indicates Dular has greater roomto maneuvre its phospholipid reserve under low Pi stress. Allthese results together indicate the higher metabolic adjustmentsin Dular relative to PB1 which allow it to withstand low Pistress. However, a matter of future research will be whether highphospholipid content and its remobilization is a general lowPi tolerance strategy among traditional rice genotypes. Earlierreports also suggest lipid remobilization is one of the importantlow Pi tolerance mechanisms in other crops like kidney bean, oat,and sesame (Gniazdowska et al., 1999; Andersson et al., 2003,2005; Shimojima et al., 2013).

Moreover, SNPs and InDels identified in the promoter regionsbetween PB1 and Dular can modulate the differential expressionof these PSR genes between these two genotypes. This differentialexpression due to non-sense and other large effect SNPs of Dularand PB1 can provide the tolerance and sensitive attributes tothese genotypes. Further characterization of these large effectvariants will allow the detailed dissection of the molecular

Frontiers in Plant Science | www.frontiersin.org 13 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

mechanism of low P tolerance of Dular and other traditionalcultivars. However, a large population will be needed to explorethese SNPs to achieve the necessary genotypic and phenotypicvariation.

In our data, we also found induction of various transporters inDular, including the glucose six phosphate transporter, glycerolthree phosphate transporters, and phospholipid transporters,which are involved in cellular Pi remobilization. This providesadditional evidence of better internal Pi utilization in Dular.Pi transporters (both low and high-affinity) mediate uptakefrom soil and distribution of Pi in plants against concentrationgradients (Ai et al., 2009; Li et al., 2015). However, induction ofthe high-affinity Pi transporter was observed as a response ratherthan a tolerance mechanism under low Pi (Pariasca-Tanaka et al.,2009; Oono et al., 2013). High induction of OsPT6 in PB1 rootsmight reflect its greater Pi demand from exogenous sources. Onthe contrary, Dular’s higher Pi use efficiency and low Pi demandis potentially addressed by better cellular Pi mobilization andefficient Pi acquisition. This may explain the relatively lowerinduction of high affinity Pi transporters in Dular.

In conclusion, we have shown that the traditional genotypeDular might employ both better internal Pi utilization andacquisition strategies for low Pi tolerance as revealed bytranscriptome, lipidome and RSA phenotyping. Traditionalupland genotypes like Dular can serve as a novel genetic sourcesfor improving low Pi tolerance of modern Pi responsive elitecultivars.

AUTHOR NOTE

Microarray data submitted in Gene Expression Omnibus (GEO)database, www.ncbi.nih.gov/geo (accession no. GSE74795).

AUTHOR CONTRIBUTIONS

PM and BP conducted the experiments and contributed inwriting the manuscript. JG conceived the idea, designed theproject, analyzed data, and wrote the manuscript with the helpfrom co-authors.

ACKNOWLEDGMENTS

This work was supported by the research grant of DBT(Grant No. BT/PR3299/AGR/2/813/2011), Ministry ofScience and Technology, Government of India. PM andBP acknowledge the research fellowship by CSIR and DBT,respectively. We thank Prof. Jonathan Lynch and Prof. AkhileshK. Tyagi for valuable suggestions and critical readings ofmanuscript.

SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be foundonline at: http://journal.frontiersin.org/article/10.3389/fpls.2015.01184

REFERENCES

Ai, P., Sun, S., Zhao, J., Fan, X., Xin, W., Guo, Q., et al. (2009). Two rice

phosphate transporters, OsPht1; 2 and OsPht1; 6, have different functions

and kinetic properties in uptake and translocation. Plant J. 57, 798–809. doi:

10.1111/j.1365-313X.2008.03726.x

Andersson, M. X., Larsson, K. E., Tjellstrom, H., Liljenberg, C., and Sandelius, A.

S. (2005). Phosphate-limited oat. The plasma membrane and the tonoplast as

major targets for phospholipid-to-glycolipid replacement and stimulation of

phospholipases in the plasma membrane. J. Biol. Chem. 280, 27578–27586. doi:

10.1074/jbc.M503273200

Andersson, M. X., Stridh, M. H., Larsson, K. E., Liljenberg, C., and Sandelius,

A. S. (2003). Phosphate-deficient oat replaces a major portion of the plasma

membrane phospholipids with the galactolipid digalactosyldiacylglycerol. FEBS

Lett. 537, 128–132. doi: 10.1016/S0014-5793(03)00109-1

Cai, H., Xie, W., and Lian, X. (2013). Comparative analysis of differentially

expressed genes in rice under nitrogen and phosphorus starvation stress

conditions. Plant Mol. Bio. Rep. 31, 160–173. doi: 10.1007/s11105-012-0485-8

Calderon-Vazquez, C., Ibarra-Laclette, E., Caballero-Perez, J., and Herrera-

Estrella, L. (2008). Transcript profiling of Zeamays roots reveals gene responses

to phosphate deficiency at the plant- and species-specific levels. J. Exp. Bot. 59,

2479–2497. doi: 10.1093/jxb/ern115

Cao, P. J., Bartley, L. E., Jung, K. H., and Ronald, P. C. (2008). Construction of a rice

glycosyltransferase phylogenomic database and identification of rice-diverged

glycosyltransferases.Mol. Plant 1, 858–877. doi: 10.1093/mp/ssn052

Chin, J. H., Lu, X., Haefele, S. M., Gamuyao, R., Ismail, A., Wissuwa, M., et al.

(2010). Development and application of gene-based markers for the major

rice QTL Phosphorus uptake 1. Theor. Appl. Genet. 120, 1073–1086. doi:

10.1007/s00122-009-1235-7

Dai, X., Wang, Y., Yang, A., and Zhang, W. H. (2012). OsMYB2P-1, an R2R3

MYB transcription factor, is involved in the regulation of phosphate-starvation

responses and root architecture in rice. Plant Physiol. 159, 169–183. doi:

10.1104/pp.112.194217

Deveshwar, P., Bovill,W. D., Sharma, R., Able, J. A., and Kapoor, S. (2011). Analysis

of anther transcriptomes to identify genes contributing to meiosis and male

gametophyte development in rice. BMC Plant Biol. 11:78. doi: 10.1186/1471-

2229-11-78

Dormann, P. (2007). Galactolipids in plant membranes. Encyclopedia of Life

Sciences: eLS. Chichester, UK: John Wiley & Sons Ltd.

Gamuyao, R., Chin, J. H., Pariasca-Tanaka, J., Pesaresi, P., Catausan, S., Dalid, C.,

et al. (2012). The protein kinase Pstol1 from traditional rice confers tolerance

of phosphorus deficiency. Nature 488, 535–539. doi: 10.1038/nature11346

Gerloff, S. (1977). “Plant efficiencies in use of N, P and K,” in Plant Adaptation

to Mineral Stress in Problem Soils, ed M. J. Wright (New York, NY: Cornell

University Press Ithaca), 61–74.

Gilbert, N. (2009). Environment: the disappearing nutrient. Nature 461, 716–718.

doi: 10.1038/461716a

Girke, T., Lauricha, J., Tran, H., Keegstra, K., and Raikhel, N. (2004). The cell

wall navigator database. A systems-based approach to organism-unrestricted

mining of protein families involved in cell wall metabolism. Plant Physiol. 136,

3003–3008. doi: 10.1104/pp.104.049965

Gniazdowska, A., Szal, B., and Rychter, A. M. (1999). The effect of phosphate

deficiency onmembrane phospholipid composition of bean (Phaseolus vulgaris

L.) roots. Acta Physiol. Plant. 21, 263–269. doi: 10.1007/s11738-999-0041-9

Ha, S., and Tran, L. S. (2013). Understanding plant responses to phosphorus

starvation for improvement of plant tolerance to phosphorus deficiency

by biotechnological approaches. Crit. Rev. Biotechnol. 34, 16–30. doi:

10.3109/07388551.2013.783549

Hammond, J. P., Broadley, M. R., Cra igon, D. J., Higgins, J., Emmerson, Z. F.,

Townsend, H. J., et al. (2005). Using genomic DNA-based probe-selection to

improve the sensitivity of high-density oligonucleotide arrays when applied to

heterologous species. Plant Methods 1:10. doi: 10.1186/1746-4811-1-10

Frontiers in Plant Science | www.frontiersin.org 14 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

Henry, A., Gowda, V. R. P., Torres, R. O., McNally, K. L., and Serraj, R. (2011).

Variation in root system architecture and drought response in rice (Oryza

sativa): phenotyping of the OryzaSNP panel in rainfed lowland fields. Field

Crops Res. 120, 205–214. doi: 10.1016/j.fcr.2010.10.003

Hernández, G., Ramírez, M., Valdés-López, O., Tesfaye, M., Graham, M.

A., Czechowski, T., et al. (2007). Phosphorus stress in common bean:

root transcript and metabolic responses. Plant Physiol. 144, 752–767. doi:

10.1104/pp.107.096958

Hill, J. O., Simpson, R. J., Moore, A. D., and Chapman, D. F. (2006). Morphology

and response of roots of pasture species to phosphorus and nitrogen nutrition.

Plant Soil 286, 7–19. doi: 10.1007/s11104-006-0014-3

Jain, A., Poling, M. D., Karthikeyan, A. S., Blakeslee, J. J., Peer, W. A.,

Titapiwatanakun, B., et al. (2007). Differential effects of sucrose and auxin

on localized phosphate deficiency-induced modulation of different traits of

root system architecture in Arabidopsis. Plant Physiol. 144, 232–247. doi:

10.1104/pp.106.092130

Kirk, G. J. D., and Du, L. V. (1997). Changes in rice root architecture, porosity,

and oxygen and proton release under phosphate deficiency. New Phytol. 135,

191–200. doi: 10.1046/j.1469-8137.1997.00640.x

Li, L., Liu, C., and Lian, X. (2010). Gene expression profiles in rice roots under

low phosphorus stress. Plant Mol. Biol. 72, 423–432. doi: 10.1007/s11103-009-

9580-0

Li, Y., Zhang, J., Zhang, X., Fan, H., Gu, M., Qu, H., et al. (2015).

Phosphate transporter OsPht1;8 in rice plays an important role in phosphorus

redistribution from source to sink organs and allocation between embryo and

endosperm of seeds. Plant Sci. 230, 23–32. doi: 10.1016/j.plantsci.2014.10.001

López-Arredondo, D. L., Leyva-González, M. A., González-Morales, S. I., López-

Bucio, J., and Herrera-Estrella, L. (2014). Phosphate nutrition: improving

low-phosphate tolerance in crops. Annu. Rev. of Plant Biol. 65, 95–123. doi:

10.1146/annurev-arplant-050213-035949

Lynch, J. P. (2011). Root phenes for enhanced soil exploration and phosphorus

acquisition: tools for future crops. Plant Physiol. 156, 1041–1049. doi:

10.1104/pp.111.175414

Matyash, V., Liebisch, G., Kurzchalia, T. V., Shevchenko, A., and Schwudke,

D. (2008). Lipid extraction by methyl-tert-butyl ether for high-throughput

lipidomics. J. Lipid Res. 49, 1137–1146. doi: 10.1194/jlr.D700041-JLR200

Mehra, P., Pandey, B. K., and Giri, J. (2015). Genome-wide DNA polymorphism

in low Phosphate tolerant and sensitive rice genotypes. Sci. Rep. 5:13090. doi:

10.1038/srep13090

Misson, J., Raghothama, K. G., Jain, A., Jouhet, J., Block, M. A., Bligny, R., et al.

(2005). A genome-wide transcriptional analysis using Arabidopsis thaliana

Affymetrix gene chips determined plant responses to phosphate deprivation.

Proc. Natl. Acad. Sci. U.S.A. 102, 11934–11939. doi: 10.1073/pnas.0505266102

Müller, R., Morant, M., Jarmer, H., Nilsson, L., and Nielsen, T. H. (2007).

Genome-wide analysis of the Arabidopsis leaf transcriptome reveals interaction

of phosphate and sugar metabolism. Plant Physiol. 143, 156–171. doi:

10.1104/pp.106.090167

Mutum, R. D., Balyan, S. C., Kansal, S., Agarwal, P., Kumar, S., Kumar, M.,

et al. (2013). Evolution of variety−specific regulatory schema for expression

of osa−miR408 in indica rice varieties under drought stress. FEBS J. 280,

1717–1730. doi: 10.1111/febs.12186

Nakamura, Y. (2013). Phosphate starvation and membrane lipid remodeling in

seed plants. Prog. Lipid Res. 52, 43–50. doi: 10.1016/j.plipres.2012.07.002

Okazaki, Y., Otsuki, H., Narisawa, T., Kobayashi, M., Sawai, S., Kamide, Y., et al.

(2013). A new class of plant lipid is essential for protection against phosphorus

depletion. Nat. commun. 4, 1510. doi: 10.1038/ncomms2512

Oono, Y., Kawahara, Y., Yazawa, T., Kanamori, H., Kuramata, M., Yamagata, H.,

et al. (2013). Diversity in the complexity of phosphate starvation transcriptomes

among rice cultivars based on RNA-Seq profiles. Plant Mol. Biol. 83, 523–537.

doi: 10.1007/s11103-013-0106-4

Pandey, B. K., Mehra, P., and Giri, J. (2013). “Phosphorus starvation response in

plants and opportunities for crop improvement,” in Climate Change and Plant

Abiotic Stress Tolerance, eds N. Tuteja and S. S. Gill (Weinheim: Wiley-VCH

Verlag GmbH & Co. KGaA), 991–1012.

Pant, B. D., Burgos, A., Pant, P., Cuadros-Inostroza, A., Willmitzer, L., and

Scheible, W. R. (2015). The transcription factor PHR1 regulates lipid

remodeling and triacylglycerol accumulation in Arabidopsis thaliana during

phosphorus starvation. J. Exp. Bot. 66, 1907–1918. doi: 10.1093/jxb/eru535

Pariasca-Tanaka, J., Satoh, K., Rose, T., Mauleon, R., and Wissuwa, M. (2009).

Stress response versus stress tolerance: a transcriptome analysis of two rice

lines contrasting in tolerance to phosphorus deficiency. Rice 2, 167–185. doi:

10.1007/s12284-009-9032-0

Park, M. R., Baek, S. H., de Los Reyes, B. G., Yun, S. J., and Hasenstein, K. H.

(2012). Transcriptome profiling characterizes phosphate deficiency effects on

carbohydrate metabolism in rice leaves. J. Plant Physiol. 169, 193–205. doi:

10.1016/j.jplph.2011.09.002

Plaxton, W. C., and Tran, H. T. (2011). Metabolic adaptations of phosphate-

starved plants. Plant Physiol. 156, 1006–1015. doi: 10.1104/pp.111.175281

Poirier, Y., Thoma, S., Somerville, C., and Schiefelbein, J. (1991). Mutant of

Arabidopsis deficient in xylem loading of phosphate. Plant Physiol. 97,

1087–1093. doi: 10.1104/pp.97.3.1087

Postma, J. A., and Lynch, J. P. (2013). Theoretical evidence for the functional

benefit of root cortical aerenchyma in soils with low phosphorus availability.

Ann. Bot. 107, 829–841. doi: 10.1093/aob/mcq199

Raghothama, K. G. (1999). Phosphate acquisition. Annu. Rev. Plant Phys. 50,

665–693. doi: 10.1146/annurev.arplant.50.1.665

Rose, T. J., Rose, M. T., Pariasca-Tanaka, J., Heuer, S., andWissuwa, M. (2011). The

frustration with utilization: why have improvements in internal phosphorus

utilization efficiency in crops remained so elusive? Front. Plant Sci. 2:73. doi:

10.3389/fpls.2011.00073

Rouached, H., Arpat, A. B., and Poirier, Y. (2010). Regulation of phosphate

starvation responses in plants: signaling players and cross-talks. Mol. Plant 3,

288–299. doi: 10.1093/mp/ssp120

Secco, D., Jabnoune, M., Walker, H., Shou, H., Wu, P., Poirier, Y., et al.

(2013). Spatio-temporal transcript profiling of rice roots and shoots in

response to phosphate starvation and recovery. Plant Cell 25, 4285–4304. doi:

10.1105/tpc.113.117325

Sharma, D., Sagwal, P. K., Singh, I., and Sangwan, A. (2012). Influence of different

nitrogen and phosphorus levels on profitability, plant nutrient content, yield

and quality in basmati cultivars. IJASER, 1, 1–4. doi: 10.4236/ajps.2015.

69135

Shimizu, A., Yanagihara, S., Kawasaki, S., and Ikehashi, H. (2004). Phosphorus

deficiency-induced root elongation and its QTL in rice (Oryza sativa L.). Theor.

Appl. Genet. 109, 1361–1368. doi: 10.1007/s00122-004-1751-4

Shimojima, M., Watanabe, T., Madoka, Y., Koizumi, R., Yamamoto, M.

P., Masuda, K., et al. (2013). Differential regulation of two types of

monogalactosyldiacylglycerol synthase in membrane lipid remodeling under

phosphate-limited conditions in sesame plants. Front. Plant Sci. 4:469. doi:

10.3389/fpls.2013.00469

Svistoonoff, S., Creff, A., Reymond, M., Sigoillot-Claude, C., Ricaud, L., Blanchet,

A., et al. (2007). Root tip contact with low-phosphate media reprograms plant

root architecture. Nat. Genet. 39, 792–796. doi: 10.1038/ng2041

Takehisa, H., Sato, Y., Igarashi, M., Abiko, T., Antonio, B. A., Kamatsuki,

K., et al. (2012). Genome-wide transcriptome dissection of the rice root

system: implications for developmental and physiological functions. Plant J. 69,

126–140. doi: 10.1111/j.1365-313X.2011.04777.x

Vissenberg, K., Fry, S. C., and Verbelen, J. P. (2001). Root hair initiation is coupled

to a highly localized increase of xyloglucan endotransglycosylase action in

Arabidopsis roots. Plant Physiol. 127, 1125–1135. doi: 10.1104/pp.010295

Wang, L., Li, Z., Qian, W., Guo, W., Gao, X., Huang, L., et al. (2011). The

Arabidopsis purple acid phosphatase AtPAP10 is predominantly associated with

the root surface and plays an important role in plant tolerance to phosphate

limitation. Plant Physiol. 157, 1283–1299. doi: 10.1104/pp.111.183723

Wang, X., Shen, J., and Liao, H. (2010). Acquisition or utilization, which is more

critical for enhancing phosphorus efficiency in modern crops? Plant Sci. 179,

302–306. doi: 10.1016/j.plantsci.2010.06.007

Wang, Y. H., Garvin, D. F., and Kochian, L. V. (2002). Rapid induction of

regulatory and transporter genes in response to phosphorus, potassium, and

iron deficiencies in tomato roots. Evidence for cross talk and root/rhizosphere-

mediated signals. Plant Physiol. 130, 1361–1370. doi: 10.1104/pp.008854

Wasaki, J., Shinano, T., Onishi, K., Yonetani, R., Yazaki, J., Fujii, F., et al.

(2006). Transcriptomic analysis indicates putative metabolic changes caused

by manipulation of phosphorus availability in rice leaves. J. Exp. Bot. 57,

2049–2059. doi: 10.1093/jxb/erj158

Wasaki, J., Yonetani, R., Kuroda, S., Shinano, T., Yazaki, J., Fujii, F., et al.

(2003). Transcriptomic analysis of metabolic changes by phosphorus stress

Frontiers in Plant Science | www.frontiersin.org 15 January 2016 | Volume 6 | Article 1184

Mehra et al. Pi-Efficient Strategies of a Traditional Rice Genotype

in rice plant roots. Plant Cell Environ. 26, 1515–1523. doi: 10.1046/j.1365-

3040.2003.01074.x

Wissuwa, M., and Ae, N. (2001). Genotypic variation for tolerance to phosphorus

deficiency in rice and the potential for its exploitation in rice improvement.

Plant Breeding 120, 43–48. doi: 10.1046/j.1439-0523.2001.00561.x

York, L. M., Nord, E. A., and Lynch, J. P. (2013). Integration of root phenes for soil

resource acquisition. Front. Plant Sci. 4:355. doi: 10.3389/fpls.2013.00355

Yoshida, S., Forno, D. A., Cock, J. H., and Gomez, K. A. (1976). LaboratoryManual

for Physiological Studies of Rice, Edn. 3. Manila: International Rice Research

Institute.

Conflict of Interest Statement: The authors declare that the research was

conducted in the absence of any commercial or financial relationships that could

be construed as a potential conflict of interest.

Copyright © 2016 Mehra, Pandey and Giri. This is an open-access article distributed

under the terms of the Creative Commons Attribution License (CC BY). The use,

distribution or reproduction in other forums is permitted, provided the original

author(s) or licensor are credited and that the original publication in this journal

is cited, in accordance with accepted academic practice. No use, distribution or

reproduction is permitted which does not comply with these terms.

Frontiers in Plant Science | www.frontiersin.org 16 January 2016 | Volume 6 | Article 1184